Using AOD and UVAI to Reduce the Uncertainties in Wildfire Emission and Air Quality Modeling

Abstract Wildfires are a major natural source of atmospheric aerosols, leading to air quality degradation and adverse human health effects. Accurate prediction of air quality effects from wildfires remains challenging due to uncertainties in fire emission estimates. To enhance the accuracy of fire emissions used in air quality forecast models, we developed a method that utilizes satellite aerosol optical depth (AOD) observations and air quality simulations to calculate dynamic emission scaling factors and improve wildfire air quality forecasts. TROPOMI (TROPOspheric Monitoring Instrument) UV Aerosol Index (UVAI) data are employed to fill AOD gaps under thick smoke using two approaches: a regression model and an artificial intelligence model. The scaling factor method was applied to NOAA blended Global Biomass Burning Emissions Product. The emission scaling factors exhibited significant variability across different fire points, highlighting the need for point‐specific scaling factors. On average, scaling factors were less than 1.0 (indicating emission overestimation) during the initial stages of fire events but exceeded 1.0 (suggesting underestimation) after 7 days of fire duration. An inverse relationship between scaling factors and fire radiative power (FRP) was observed, with emission underestimation for low‐intensity fires (FRP <5 MW) and substantial overestimation for high‐intensity fires (FRP >500 MW). The improved fire emissions were employed in the air quality model for the 2020 US Gigafire event. Utilizing emission scaling factors reduced model bias, increased the correlation and hit rate of PM 2.5 exceedance prediction, demonstrating the potential of using emission scaling factors for improving air quality forecasting during wildfire events.